最近(최근) chardet의 MIT 리라이선싱과 關聯(관련)해서 Redis의 antirez와 Flask의 Armin Ronacher(一名(일명) mitsuhiko)가 이를 擁護(옹호)하는 글을 썼는데, 그에 對(대)한 나름대로의 反論(반론)을 써 봤습니다: 〈合法(합법)이면 公正(공정)한가: AI 再具顯(재구현)과 카피레프트의 侵蝕(침식)〉(한글).

@hongminhee The moral ambit of LLM users can be concluded with the sentence "there is an obvious moral question here, but that isn't necessarily what I'm interested in" from what I've gathered.

I've seen people getting excited by what they perceive as copyright being no longer applicable because if companies can ingest FOSS code and make it proprietary, the commons can apparently do the same and that seems to miss the asymmetry of power. An individual like Aaron Swartz can still be mentally harassed to the point where he kills himself for violating copyright in the age of LLMs but there will never be equivalent consequences for Meta for pirating TBs of books.

@hongminhee @mitsuhiko @antirez
I would also note that, currently, the extension of copyright protection to AI-generated works remains an open question, and the US side of the common law does not consider them even copyrightable without a significative human contribution (no, prompts do not count). So, every shot to the copyleft has its consequences I would say. The whole matter is quite complicated and questionable.

@hongminhee @mitsuhiko @antirez
It is technically legal, but note that it took ages to create a legal framework fully accepted for a lot of products of creativity that were fully accepted by all parties. And still there were corners in the rights not fully understood.

Even, note that the white room reimplementation has been only a technical trick to enforce the safety of a reimplementation, not a strict legal requirement.

For the future, who knows?